INVITED ARTICLE

What’s Next?
After Stage-Gate
Progressive companies are developing a new generation of idea-to-launch processes.

Robert G. Cooper

OVERVIEW: As the creator of the Stage-Gate® process, I am often asked, “What’s next after stage-gate?” For years, I’ve not
had an answer. Now, we’re seeing new approaches emerging from progressive companies that represent a new generation
of idea-to-launch processes. In some cases, it’s an evolution of Stage-Gate to a better, faster model; in other firms, its closer
to a revolution, moving to a very different system. But there is anything but unanimity as to what the next generation idea-
to-launch system should be. This article looks at what leading firms are doing to move beyond their current idea-to-launch
methodology and tries to integrate these practices into a next-generation system.
KEYWORDS: Stage-Gate, Phase review, Gating, Idea-to-launch process, Agile development, Accelerated development,
Adaptive development

The original Stage-Gate system was created in the 1980s, Edgett 2012). But there are also criticisms, some the result
based on an in-depth study of successful “intrapreneurs” of the nature of the process and others of the way compa-
within major corporations as they drove successful new nies implemented the system. The world has changed a lot
products to market. Their practices and the lessons they since the first Stage-Gate system was implemented—it is
learned provided the foundation for that early stage-and- now faster paced, more competitive and global, and less
gate model. Over the years, Stage-Gate has evolved and in- predictable. In this context, Stage-Gate has attracted a num-
corporated many new practices (see, for instance, Cooper ber of criticisms: It is accused of being too linear, too rigid,
1994, 2008, 2011). Some companies have also developed and too planned to handle more innovative or dynamic
their own versions of Stage-Gate, building in some positive projects. It’s not adaptive enough and does not encourage
elements, but also some negative ones. experimentation. It’s not context-based—one size should
Today we see that the Stage-Gate process has generally not fit all. Its gates are too structured or too financially
had a positive impact on the conception, development, and based, and the system is too controlling and bureaucratic,
launch of new products (Cooper 2011, 2013a; Cooper and loaded with paperwork, checklists, and too much non-
value-added work (Becker 2006; Lenfle and Loch 2010).
Some authors have taken issue with these criticisms, argu-
Robert G. Cooper is president of the Product Development Institute. He is ing that most are due to faulty implementation (Becker
a professor emeritus at McMaster University’s DeGroote School of Business, 2006), while some deficiencies have been corrected in more
ISBM Distinguished Research Fellow at Penn State University’s Smeal Col-
recent evolutions of Stage-Gate (Cooper 2011). But issues
lege of Business Administration, and a Crawford Fellow of the Product De-
velopment and Management Association. A thought leader in the field of do remain, and thus a handful of leading firms are rethink-
product innovation management and developer of the Stage-Gate new ing and re-inventing their idea-to-launch gating system.
product development process, he has won two IRI Maurice Holland awards Through my ongoing study of benchmarking best practices,
and has published over 120 articles and seven books. He received his PhD
in business administration from the University of Western Ontario and Bach-
presentations at the annual Stage-Gate Innovation Summit,
elors and Masters degrees in chemical engineering from McGill University. and personal interactions with leading firms, I’ve con-
robertcooper@cogeco.ca structed an overview of likely directions for the next gen-
DOI: 10.5437/08956308X5606963 eration of idea-to-launch systems.

20 | Research-Technology Management • January—February 2014

The Next Generation Idea-to-Launch System: The Triple criteria for each gate—and gates are integrated with portfo-
A System lio management.
At first glance, the practices and recommendations of firms A2—Agile: The next-generation system also incorpo-
creating new idea-to-launch systems look a lot like the tradi- rates elements of Agile Development, the rapid develop-
tional process; there are still stages where work gets done, ment system developed by the software industry. For
and there are still gates where decisions are made. But the example, sprints and scrums—short time-boxed incre-
details of the process and its function are quite different: ments in which the deliverable is something that can be
What emerges is a more agile, vibrant, dynamic, flexible gat- demonstrated to stakeholders (rather than documenta-
ing process that is leaner, faster, and more adaptive and risk- tion)—are part of the new system. Equally, these new sys-
based. This is what I call the Triple A system—it is adaptive tems emphasize moving quickly and nimbly from
and flexible, agile, and accelerated (Figure 1). milestone to milestone and rely on a much leaner system
A1—Adaptive and Flexible: The next-generation with all waste removed—no bureaucracy, no unnecessary
idea-to-launch system is adaptive. It incorporates spiral or activities anywhere in the system.
iterative development to get something in front of custom- A3—Accelerated: The next-generation idea-to-launch
ers early and often through a series of build-test-revise it- system is focused on accelerating the development pro-
erations. The product may be less than 50 percent defined cess. Projects in the system are properly resourced, espe-
when it enters development, but it evolves, adapting to new cially major projects, and fully staffed by a dedicated
information, as it moves through development and testing. cross-functional team for maximum speed to market. Ac-
The system is also flexible insofar as the actions for each tivities within stages overlap, and even stages overlap; the
stage and the deliverables to each gate are unique to each notion of a “stage” is less relevant in this new system.
development project, based on the context of the market There is more emphasis on the fuzzy front end, making it
and the needs of the development process. This is the op- sharper and less fuzzy, so that the project is clearly scoped
posite of an SOP (standard operating procedure) approach and key unknowns, risks, and uncertainties identified as
to product development, which prescribes standardized ac- early as possible. Finally, robust IT support is provided to
tions and deliverables. There are also fast-track versions of reduce work, provide better communication, and acceler-
the process for lower-risk projects. And in the next-generation ate the process.
system, a risk-based contingency model dictates that appro-
priate activities and deliverables be determined based on an An Adaptive and Flexible Process
assessment of project assumptions and risks. Finally, Go/Kill Emerging idea-to-launch systems take their power from
criteria are flexible—there are no standard sets or universal being adaptive and flexible—able to shape themselves to

FIGURE 1. The next-generation idea-to-launch system

What’s Next? After Stage-Gate January—February 2014 | 21

the context of particular projects. These qualities arise much more adaptive. The product may be less than 50 per-
from four attributes in particular: spiral development cent defined as development begins, but comes together
cycles, context-based stage definitions and activities, risk- during development; the product’s design and definition
based contingency models to drive decision making, and adapts to new information, customer feedback, and chang-
flexible criteria for Go/Kill decisions. ing conditions on its way to launch. Such firms have built
in multiple spirals or iterations of development that permit
Spiral Development experimentation with users (Figure 2), with each spiral
The traditional Stage-Gate process requires that the product consisting of:
and project be defined before the project moves into devel-
• Build. In each iteration, build something to show the
opment. Indeed, “sharp, early and fact-based product defi-
customer—a rapid prototype, a protocept, a crude work-
nition” is a fundamental principle of Stage-Gate (Cooper
ing model, an early beta version.
2011, 2013a). But the world moves too fast today to make
• Test. Test each version of the product with customers—let
a stable and rigid product definition possible for some busi-
them tell you what they like and what value they see.
nesses and projects. Often customers are not clear on what
• Feedback. Gather feedback on that version of the product
they want (or need) in the first place, so it’s impossible to
from the customer or user.
get a 100 percent accurate product definition prior to devel-
• Revise. Reset your thinking about the value proposition,
opment. As Steve Jobs, never a proponent of traditional
benefits sought, and the product’s design based on the
market research, famously said, “People don’t know what
feedback, and start again.
they want until you show it to them” (Isaacson 2011, 567).
And sometimes requirements simply change in the time Each loop moves the project closer to the final product
that passes between the beginning and end of development— design. This spiral approach promotes experimentation,
a new customer need, a new competitive product, or a new encouraging project teams to fail often, fail fast, and fail
technological possibility emerge, and the original product cheaply, a principle that Jobs applied throughout his devel-
definition is no longer valid. opment career at Apple (Isaacson 2011). Corning has ad-
Thus smart firms, especially those doing riskier and opted this approach with 60–90-day plans that include
bolder projects, have made the idea-to-launch system numerous iterations that yield testable versions of the

FIGURE 2. Spiral development built on build-test-feedback-revise iterations

22 | Research-Technology Management What’s Next? After Stage-Gate

product as a deliverable at key milestones (often weekly or • An express version for very small developments, such as
biweekly). This looping or spiral development is consistent a sales-force request requiring a minor product change.
with two core tenets of the Agile Manifesto for software
In addition, many firms have introduced special versions
development: a focus on quick response to change, and
of Stage-Gate to handle technology development projects
continuous customer or stakeholder involvement in the de-
where the deliverable is not a new product, but rather a
velopment of the product.
technological capability (Cooper 2006). Such projects
still require the discipline of a gating process, but the ac-
Context-Based Stages and Gates tivities, deliverables, and criteria for Go are unique to
In the first generations of Stage-Gate, a single model was de- these projects; hence, they merit their own system. Some
veloped, intended to handle the most difficult or complex new examples:
product projects within a company. The irony is that most
• The Kellogg Company has three versions of its K-Way
projects the company undertook were much simpler, and only
innovation process: the regular five-stage process to
a handful were the complex initiatives that the process had
handle new products, a lighter three-stage process for
been designed for. Many firms have now developed faster-
smaller projects, and a three-stage process to handle
track or lighter versions of Stage-Gate to handle less risky,
technology developments such as new science or in-
better-defined, and less complex development projects; recent
vention projects (Maley 2010).
benchmarking studies show that 75 percent of top-performing
• 3M has its regular new product innovation (NPI) system,
businesses use a scalable idea-to-launch process (Cooper and
a standard five-stage model to handle typical new-product
Edgett 2012). For example, there are often three versions of
projects. There is also a shorter three-stage, three-gate
Stage-Gate (Figure 3):
version used by international subsidiaries, where the US-
• The full five-stage process to handle major, high-risk developed product is modified for sale locally. Finally,
developments; there is the three-stage new technology innovation (NTI)
• A light version for moderate risk projects, such as signifi- system for managing the design and development of a
cant modifications, improvements, and extensions; and new technology (Gehring 2011).

FIGURE 3. Scalable Stage-Gate systems

What’s Next? After Stage-Gate January—February 2014 | 23

• P&G, by contrast, has decided not to employ different pro- • The traditional phase-review system to deliver lower
cesses for fast-track projects versus its durable five-stage costs or new features for mature markets, such as desk-
SIMPL process. As Dietmar Bressau, Corporate Leader for top PCs (MacCormack et al. 2012).
Innovation Diamond Management, explained in an inter-
view, “When you open the door to two to three processes,
The Risk-Based Contingency Model
then along comes four, five, and six. Ultimately, anything
Perhaps the most significant departure from standard gating
goes. We want to keep one common Stage-Gate process
systems is the ability to custom tailor the process to each
that will be tailored based on the risk profile of an indi-
and every project—the polar opposite of an SOP approach.
vidual initiative—otherwise each team, functional area, or
In this approach, the entire new product process, from idea
business does its own thing.” Nonetheless, the firm has
to launch, is viewed as a series of steps and activities de-
developed a value-driven process for innovation that re-
signed to gather information to reduce uncertainty and
quires invention. It’s called FEI, or Front End Innovation;
thereby manage risk (Cooper 2011, 94–98). Thus the na-
the argument is that when one brings a project into SIMPL
ture of the uncertainties and risk specific to a project should
without the invention in place, the project spends forever
determine what happens in that project. But note that in-
in the design phase. Thus, the technology development
formation only has a value to the extent that it can improve
must be handled in advance.
a decision (reduce project uncertainties or validate assump-
• Hewlett-Packard recognized that its traditional phase-
tions) that has economic consequences.
review process, while excellent for mature markets re-
Thus, the project team begins with a blank canvas, then
quiring product improvements and extensions, was not
so well suited to emerging, fast-moving markets. To meet
• Identifies key unknowns and uncertainties,
the needs of these different types of markets, HP now
• Pinpoints the critical assumptions (critical in the sense
defines three development processes:
that the assumptions have economic consequences),
• An emergent model for start-up developments, such as and
cloud computing; • Determines what information is needed to validate these
• An agile model for growth sectors, such as blade servers; and assumptions.

FIGURE 4. Corning’s risk-based contingency model (adapted from Kirk 2013)

24 | Research-Technology Management What’s Next? After Stage-Gate

The information needs in turn define the deliverables re-
quired at the next project review, and hence define the ac-
tivities required in that stage. In this way, the project team Leading firms have integrated key
maps out its own set of deliverables and stage activities, spe-
cific to its own project. No activities are done simply because elements of the Agile Manifesto into
they are “on a list” in some manual, or deliverables prepared their Stage-Gate processes.
because there is a template to be completed.
The process is very flexible and efficient; every project has
its own custom-tailored methodology with no work included
that does not add value. The downside, however, is that proj-
ect teams must be very experienced to make this work. management became popular in the 1990s, but now leading
An example: Corning has developed and is piloting a firms have successfully integrated gates with portfolio man-
contingency model based on project risk (Figure 4). Key agement. Portfolio management has become increasingly vi-
assumptions and key unknowns are identified in each tal because of concerns arising over the mix and balance of
stage. And given these key assumptions, critical activities projects in firms’ development portfolios, and in particular,
are defined: “Here is what best practice project teams have the trend to smaller, low-risk projects and away from larger,
done, and thus what best practices apply to your project, more venturesome initiatives (see Cooper 2013b, especially
given its assumptions and risks.” So the key assumptions Figure 1). Gates occur in real time and look at individual
help to define the required activity set for the project. Ex- projects in depth; gates are often where a project in trouble is
perienced team members have no problem identifying killed after a thorough review. By contrast, portfolio reviews
what is important to do, what tasks are critical, and what occur periodically (typically four times per year) and look at
best practices to do—not every project requires every- the entire set of development projects—examining the mix,
thing. Gorilla® Glass, which is used for the screen in Ap- balance, and prioritization of projects.
ple’s iPad, was developed by such a team using this method
An Agile Process
(Kirk 2013).
A second goal of next-generation idea-to-launch systems is
to create a more agile and nimble process. Some of the prac-
Flexible Criteria for Go/Kill Decisions tices here overlap with the adaptive and flexible methods
Most gating processes rely heavily on financial criteria to outlined above, and some, but not all, are derived from Agile
make Go/Kill decisions. The extension of context-based Development for software.
models, however, requires rethinking the investment criteria The various agile methods, based on the Agile Manifesto,1
for these decisions. Research evidence shows that financial promote development, teamwork, collaboration, and process
criteria do not yield the best portfolios, not so much because adaptability throughout the life cycle of the project. Two key
the financial models are theoretically wrong, but because principles of Agile Development are:
data are missing (see the review in Cooper 2011, 264–266).
• Working software is delivered frequently (in weeks
In short, business cases are wrong, especially for the most in-
rather than months), and
novative initiatives, and often by orders of magnitude. Thus
• Working software is the principal measure of progress
progressive firms are adopting more flexible criteria.
(Beck et al. 2001).
This flexibility may be most important for technology de-
velopment projects, which must rely more on strategic cri- Agile methods thus break the development process into
teria than financial, since it is often difficult to predict the small increments with minimal planning; these increments,
longer-term economic impact of the technology. Exxon known as “sprints,” are “timeboxed,” limited to very short
Chemical was one of the first firms to develop a Stage-Gate time frames, typically from one to four weeks, followed by
system for research-based products, and the firm selected a team meeting called a “scrum” (the term is derived from
non-financial criteria (such as strategic fit, competitive ad- rugby—the scrum is similar to, although somewhat more
vantage, and market potential) as Go/Kill drivers rather physical than, American football’s huddle; Takeuchi and
than profitability metrics. Nonaka [1986] introduced the term in this context). At the
Similarly, models proposed for highly innovative proj- end of each sprint, the development team must deliver a
ects, where financial projections are likely to be unreliable, working product that has been demonstrated to stakehold-
use scorecard approaches consisting of a mixture of strate- ers. Multiple iterations might be required to release a prod-
gic, competitive, leverage, and financial criteria. Addition- uct or new features.
ally, more appropriate financial approaches are employed to Physical product development, however, is much different
deal more effectively with project risk, such as options pric- from software development. For one thing, software develop-
ing theory, Monte Carlo simulation, and Expected Com- ment is almost infinitely divisible. A software development
mercial Value (for an outline of some techniques appropriate consisting of one million lines of code can be broken down
for innovation projects, see Cooper 2013b). into one hundred increments of roughly 10,000 lines, each
Finally, the gates and their Go/Kill decisions are no longer
stand-alone. Stage-Gate was developed before portfolio 1
See “Manifesto for Software Development” at www.agilemanifesto.org.

What’s Next? After Stage-Gate January—February 2014 | 25

increment yielding a working product. But the development early stages, and to the point where something can be shown
of a new machine, a new food product, or a new polymer can- to management and a customer.
not be incrementalized in this way, and thus the notion of Here’s how it works: If such an idea creates enough excite-
short time-boxed sprints does not apply quite as well. ment in the senior management team, the project is approved
In spite of these differences, leading firms have integrated and resources are provided for a six-month period, unfettered,
key elements of the Agile Manifesto into their Stage-Gate with no rules and no reviews—in effect, a six-month sprint. At
processes. They have built in time-boxed sprints (although the end of the six months, the project team must have some-
these are longer than the one to four weeks allowed for soft- thing to show that has been seen and tested by a customer. At
ware development) for which the deliverable is something that point, there are three possibilities for the project:
physical that can be demonstrated. The emphasis is on results
1. Termination.
rather than on documentation, and projects move quickly
2. Transition to an ATOM project to develop a new product
from milestone to milestone. At each milestone, actual
following all the traditional rules and regulations.
results—for example, a validated working model—are checked
3. Continuation of the BTOM project for another six
against scheduled results.
months.
Here, too, Corning provides an example. For special
projects—large, higher-risk, bold projects, such as Gorilla The method is still quite experimental, but it has shown
Glass—Corning subdivides the development and testing promising results, moving projects ahead that normally
phases into discrete increments defined in a 60–90-day plan; would have been killed in the early stages.
these are much like sprints but they sometimes last months
rather than weeks. Within these increments, there are mul- Integrating Agile with Stage-Gate
tiple one-day meetings with senior management—some- Adopting Agile doesn’t mean abandoning Stage-Gate. Indeed,
times weekly—to review and move the project along. At the as Karlstron and Runeson (2006) note in their case study
end of each increment, there is a major milestone review at of two high-technology firms using agile and Stage-Gate sys-
which the project team must deliver something that can be tems, the Stage-Gate framework can provide important sup-
demonstrated to stakeholders; there are physical milestone port for agile development processes: “software development
objectives for milestone reviews. The project moves quickly projects are not isolated activities. They usually exist as
from management review meeting to management review sub-projects in an environment composed of hardware de-
meeting and milestone to milestone (rather than from gate to velopment, marketing, production planning etc. which all
gate, as in Corning’s traditional gating process). must be managed and coordinated concurrently. . . . [A
The method is very resource intensive: the project Stage-Gate system] gives support not only for the commu-
team is 100 percent dedicated to the project, and senior nication within the project, but also for decision makers
management is very much engaged in the many reviews. A sponsoring the project or acquiring the outcome of the
project can be killed or redirected at any milestone if impor- project” (204). Similarly, Boehm and Turner (2004), dis-
tant assumptions are not validated; it does not wait until the cussing the contrasts between plan-driven software devel-
next gate for the Go/Kill decision. opment and agile approaches at length, conclude that
Omicron, a highly innovative Austrian producer of hard- future projects will need both agility and discipline, which
ware and software for testing electrical power networks, has can be achieved by containing the agile development
also integrated some Agile principles into its idea-to-launch methodology within the gate model.
system. The firm, like Corning, has a very successful Stage- Firms are arriving at this conclusion on their own. Corn-
Gate system, called ATOM (Accelerate To Market), which ing still uses its Stage-Gate framework even for these agile
handles traditional projects well. But for breakthrough proj- projects. “After more than 20 years with Stage-Gate, people
ects, when the project is vague or it’s unknown what the are very familiar with it—it’s a common reference point, and
project will even lead to, the firm uses a different process, a good communication tool across marketing, production,
BTOM (Breakthrough To Market). The BTOM system pro- and R&D,” says Bruce Kirk, Director of Corporate Innovation
vides some breathing room for ill-defined, risky projects, giv- Effectiveness. L.M. Eriksson, maker of telecommunications
ing them space to get through the “valley of death” in the equipment (hardware and software), has also integrated
agile methods with its Stage-Gate system, relying on time-
boxed iterations with physical deliverables throughout the
development and testing stages.
Adopting Agile doesn’t mean abandoning
Lean and Nimble
Stage-Gate. Indeed, the Stage-Gate Another facet of agile development that many firms have em-
braced for physical products is the drive to make the system
framework can provide important
more lean and nimble, in keeping with Agile principles.
support for agile development. The definition of bureaucracy is “work that adds no value,”
and getting rid of such work in your idea-to-launch system
is certainly consistent with yet another Agile principle:

26 | Research-Technology Management What’s Next? After Stage-Gate

“Simplicity—the art of maximizing the amount of work not
done—is essential.”
Johnson & Johnson has revised its idea-to-launch process Often, taking simple steps, such
to eliminate bureaucracy—work that doesn’t add value—
using Lean Six Sigma methods. In the company’s Ethicon as removing the time wasters and
Surgery group, business cases that had been 30–90-page doc- blockages through value-stream
uments that took weeks to prepare are now down to 4 pages
and a lot less work (Belair 2007). analysis, cuts the time to market
P&G’s SIMPL process has been similarly trimmed as the dramatically.
company moved to much leaner gates, cutting the volume
of deliverables packages to a mere six pages (Cooper and
Mills 2005).
And at Praxair, the process manager uses lean methods
Overlapping Stages and Concurrent Activities
(value-stream analysis) to see where the problems and time
An important way to accelerate projects is simultaneous
wasters are (Spero 2009). For example, a chemical reactor
execution. Here, key activities and even entire stages over-
for R&D work is typically designed for the task, not for the
lap, allowing projects to move ahead when information is
turnaround. But it takes 3 days to do the task, and then 21
reliable and stable—rather than waiting for perfect infor-
days to turn the reactor around for the next task. A reactor
mation (Figure 5). In this way, the project can be accelerated—
designed for both turnaround and task removes a big time
development becomes a rugby game, with multiple parallel
waster in the development stage.
activities, rather than a relay race, with activities strung
out in series. In some cases, it is even acceptable to move
An Accelerated Process activities from one stage to an earlier one and in effect to
A third goal of next-generation idea-to-launch systems is to overlap stages, starting one stage before the previous stage
accelerate projects. Often, taking simple steps, such as re- is finished.
moving the time wasters and blockages through value- Toyota has long used this approach; the rule in the firm’s
stream analysis, cuts the time to market dramatically. Firms development process is to synchronize processes for simulta-
are employing a range of methods to accelerate development neous execution (Morgan 2002). That is, each subsequent
projects, including overlapping stages and concurrent activi- function maximizes the utility of the stable information
ties, dedicated teams assigned to properly resource projects, available from the previous function as it becomes available.
efforts to sharpen the fuzzy front end, and automated sys- Development teams must do the most they can with only
tems to support project management. that portion of the data that is not likely to change.

FIGURE 5. Overlapping activities and stages

What’s Next? After Stage-Gate January—February 2014 | 27

 documents, and perform other routine tasks when they have
ready access to embedded templates and best-practice con-
To my knowledge, no company has yet tent. Some systems prepopulate templates for key deliver-
ables (such as status reports, presentations, and resource
implemented every element of the next- charts) with project information recorded elsewhere in the
generation system. But some have come system. As a result, documents that previously took hours or
days to prepare can be completed in minutes.2
close.
Integrating the Evidence into A Next-Generation
Idea-to-Launch System
We’ve had a glimpse into what some companies are doing to
Properly Resourced Projects with Dedicated Teams reinvent their Stage-Gate systems, and in some cases move be-
One of the major impediments to fast delivery of new prod- yond it. Integrating these various improvements and changes—
ucts is a lack of focus and inadequate resources—spreading some evolutionary, such as fast-track versions, some more
resources too thinly across too many projects and other revolutionary, such as the risk-based contingency model—
work (Cooper 2011, 2013a). Fully dedicated teams are a produces a framework for next-generation idea-to-launch sys-
must to maximize speed. Benchmarking studies show that tems, which may be contrasted with traditional Stage-Gate
top-performing businesses in new product development are development in terms of context, system design, the role of
considerably more focused than others. They also have ded- gates, and the organization of project teams (Table 1).
icated resources for product innovation: half have dedicated The traditional process is well suited to known and tradi-
product development teams (project team members are not tional product developments, which are the majority of proj-
working on a lot of other tasks), and more than half have a ects for most firms. But the newer process is designed for more
fully dedicated product innovation group that works on innovative and bolder projects targeted at less well defined but
new products full time (Cooper 2011). growing markets and relying on newer technology with tech-
nology risks. The next-generation system is adaptive and
flexible, agile and accelerated. Projects move quickly from
Sharpening the Fuzzy Front End
milestone to milestone and engage in frequent experiments or
More thoughtful scoping—better front-end homework to
spirals, with the evolving product regularly exposed to cus-
anticipate challenges in advance and get projects requiring
tomers in a series of build-test-revise iterations. Product defini-
new technology on the right track—can do much to save
tion is far from 100 percent at the beginning of development,
time downstream. A sharper, earlier definition of project
but evolves as customers confirm the product’s value proposi-
needs and solutions can help steer downstream decision
tion through protocepts, rapid prototypes, and early beta ver-
making and clarify assumptions and risks.
sions. Stages and activities overlap. Indeed, the process is
P&G is addressing this issue in the design of their Agile In-
customized to each project via the risk-based contingency ap-
novation Management (AIM) system, instituted to build capa-
proach: starting with a blank canvas, identify the uncertainties
bilities in project leaders and bring the agility of the system to
and risks, define the critical assumptions, and define the right
the next level. AIM forces a very deliberate focus on the scop-
deliverables and activities to validate the key assumptions.
ing of the project, placing an emphasis on the front-end of the
Gates are still part of the next-generation system, but they
SIMPL process and making the fuzzy front end a little less
are less relevant than in the traditional process, and they are
fuzzy. This is accomplished through a set of questions that
integrated with portfolio management and portfolio reviews.
must have clear answers before the project moves forward:
Go/Kill and prioritization decisions may be made at any of
these points, as well as at milestones, for example, if an impor-
• What is the risk of the project and what are the true re-
tant assumption is not confirmed. Go/Kill criteria are less fi-
quirements for success?
nancially focused, emphasizing more strategic, competitive,
• What do you want and when?
and leverage factors; when the criteria are financial, they em-
• What is possible and when?
ploy more appropriate financial models to account for risk and
• Do you need invention?
options buying. Senior management is more engaged in the
If the answer to the last question is yes, the project moves investment decision process, ready to commit the necessary
into P&G’s technology development system (FEI). resources to execute the project in an accelerated fashion.
Organizationally, the next-generation system requires
Automated Systems dedicated cross-functional project teams with the resources
An increasingly popular way of accelerating projects is the needed to move the project forward quickly—dedicated
use of automation software. The benefits of automation are people for important projects, not spread over multiple proj-
two-fold: less time is required to complete stage activities and ects and other tasks. Functional managers become resource
deliverables, and the administrative load associated with pro-
cess execution is reduced. For example, project team mem- 2
A number of excellent and accredited IT solutions are available to support
bers can more easily create gate deliverables, search for idea-to-launch systems. See www.stage-gate.com/certification_directory.php.

28 | Research-Technology Management What’s Next? After Stage-Gate

TABLE 1. The next-generation idea-to-launch framework compared to Stage-Gate development
Traditional Stage-Gate Next-Generation System
The Context
The Situation Mature market Existing and rapidly growing market
Market and customers well known Large potential market
Well-known technology Many market uncertainties and unknowns, some
Few market or technical uncertainties or risks technical risks
Product Type New item in a product line Innovation: bigger, bolder, more venturesome
Product improvement, modification, project
renovation Higher-risk initiative
Example New model of an HP ink-jet printer Corning Gorilla® Glass
Customer Needs Well known and stable over time Some known, some unarticulated
(via traditional market research) Many unsolved customer problems and
unresolved needs
Market Size Large and defined Existing, not necessarily large, but growing
May be maturing, flat, or even declining rapidly
Growth phase of product life cycle
Competition Red Ocean Some early competitors, who may not yet have
Many capable competitors the right or dominant solution
Undifferentiated products, even tending Differentiated products
to commoditization
Technology Maturity Mature Newer technology, but largely existing
Well known May be new to company, but familiar
In house
Technical Risks Few risks; technical hurdles can be Some risks and technical hurdles; hurdles can
overcome easily likely be overcome
Can envision a solution Technical solution largely envisioned
The Idea-to-Launch System
System Well-defined traditional Stage-Gate Triple A: adaptive & flexible, agile, accelerated
process gated system
Like a Standard Operating Procedure Context-based (every project is unique)
Stages laid out in a linear fashion Risk-based contingency approach; risks and
Activities pre-specified for each stage assumptions define deliverables and activities
Standard deliverables defined Frequent design-build-test iterations (spirals)—
(with templates) for each gate more experimental
Fairly repeatable and predictable stage activities: Nimble and quick, project moves milestone to
planned, not iterative milestone (sprints)
Clear, consistent exit and Go/Kill criteria, Overlapping stages
rigorous gate reviews, monitor to plan
Product Definition Established in detail up front (>90% specified) Partly established up front (40%–70% specified)
Based on well-understood customer Updated as feedback dictates (via design-
requirements, defined technical solutions build-test spirals)
Understanding Market and Customer Traditional market research, i.e., Voice Work with real customers who represent market
of Customer, site visits, ethnography
Determine needs, wants, desires Voice of Customer early to determine needs,
problems, market potential
Interact with users via rapid prototypes, early
beta versions, etc.
Building the Right Product Follow dominant design adding Define product, technical choices, features
customer-valued features (visible through early and rapid prototypes and beta
improvements) products to customers
Emphasize process- and cost-focused innovations Seek customer confirmation of design, value,
sales volume through spirals
The Gates
Flow Project moves gate to gate Project-driven milestone to milestone
Stages and gates very relevant Go/Kill decisions at milestones, gates, and
portfolio reviews
Stages and gates less relevant

continued

What’s Next? After Stage-Gate January—February 2014 | 29

TABLE 1. continued
Traditional Stage-Gate Next-Generation System
Go/Kill Criteria Largely financial—NPV, IRR, payback More qualitative and strategic
period Based on scorecard with both nonfinancial and
Financial risk vs. return assessment financial criteria
ECV, options pricing, and Monte Carlo to deal
with risk and options
Gatekeepers Senior & middle management from the Senior and middle management from the
business unit business unit
Sponsors must be senior to ensure adequate
resource commitments
Emphasis Gate deliverables mostly information Focus on results, not documentation
and documentation (may overemphasize reports, Being able to show something to stakeholders
often voluminous)
Deliverables Required Well-defined list of deliverables for Deliverables templates exist, but only as a
each gate (standard templates) guide
Quite disciplined Deliverables determined by project context
How Deliverables Are Defined Broad standard list for every project Team and gatekeepers identify key
assumptions and unknowns, then define key
deliverables
Project-specific deliverables
Portfolio Management Most Go/Kill decisions made at gate Gates integrated with portfolio reviews
meetings Go/Kill decisions made at both
Portfolio reviews are a check
The Organization
Organizational Structure Organized by specialized functions or as Organized as an accountable cross-functional
cross-functional project team with team project team
members drawn from functional areas Team may be a venture team and/or operate
outside the organizational structure, e.g., a
Skunk Works or off site
Resources Team members on multiple projects Team members dedicated to project
concurrently, projects often understaffed Projects properly resourced for maximum effort
Time and resource allocation decisions and focus
made at gates
Team Structure Balanced matrix: Project matrix:
Project leader is assigned to oversee Project leader is assigned to lead and oversee
project project, has control over resources
Team members assigned from functional departments
Relationship to Functional Departments Project leader shares responsibility and Project leader has full responsibility and
authority with functional managers authority for project
Joint approval and direction Functional managers assign personnel, provide
technical and marketing expertise

providers. The project team may be organized as a project ma- Mellor, S., Schwaber, K., Sutherland, J., and Thomas, D.
trix or venture team, or may even operate outside the official 2001. Principles behind the agile manifesto. Manifesto for
bureaucracy of the company (for example, as a Skunk Works). Agile Software Development. http://www.agilemanifesto.org/
To my knowledge, no company has yet implemented ev- principles.html
ery element of the next-generation system described here. Becker, B. 2006. Rethinking the Stage-Gate process—A reply
to the critics. Management Roundtable, July 12.
But some have come close. Private discussions with execu-
Belair, G. 2007. Beyond gates: Building the right NPD organiza-
tives in these firms reveal dramatically positive results. tion. First International Stage-Gate Leadership Summit, St.
So perhaps it’s time to rethink your idea-to-launch sys- Petersburg Beach, FL, Feb 20–21.
tem, borrow some of the methods outlined in this article, and Boehm, B., and Turner, R. 2004. Balancing Agility and Discipline.
strive for more a more adaptive, agile, and accelerated stage- New York: Addison Wesley.
and-gate system. Cooper, R. G. 1994. Third-generation new product pro-
cesses. Journal of Product Innovation Management 11(1):
References 3–14.
Beck, K., Beedle, M., van Bennekum, A., Cockburn, A., Cooper, R. G. 2006. Managing technology development proj-
Cunningham, W., Fowler, M., Grenning, J., Highsmith, J., ects: Different than traditional development projects.
Hunt, A., Jeffries, R., Kern, J., Marick, B., Martin, R. C., Research-Technology Management 49(6): 23–31.

30 | Research-Technology Management What’s Next? After Stage-Gate

Cooper, R. G. 2008. The Stage-Gate idea-to-launch process— Karlstrom, D., and Runeson, P. 2006. Integrating agile software
Update: what’s new and next-gen systems. Journal of Product development into Stage-Gate managed product develop-
Innovation Management 25(3): 213–232. ment. Empirical Software Engineering 11: 203–225.
Cooper, R. G. 2011. Winning at New Products: Creating Value Kirk, B. 2013. Accelerating time to market: Using a next genera-
Through Innovation, 4th ed. New York: Basic Books-Perseus. tion innovation framework. Stage-Gate Innovation SUMMIT,
Cooper, R. G. 2013a. New products: What separates the win- Miami, FL, February 26–28.
ners from the losers and what drives success. In PDMA Lenfle, S., and Loch, C. 2010. Lost roots: How project manage-
Handbook of New Product Development, 3rd ed., ed. K. B. Kahn, ment came to emphasize control over flexibility and nov-
3–34. Hoboken, NJ: John Wiley & Sons. elty. California Management Review 53(1): 32–55.
Cooper, R. G. 2013b. Where are all the breakthrough new MacCormack, A., Crandall, W., Henderson, P., and Toft, P.
products? Using portfolio management to boost innovation. 2012. Do you need a new-product development strategy?
Research-Technology Management 56(5): 25–33. Research-Technology Management 55(1): 34–43.
Cooper, R. G., and Edgett, S J. 2012. Best practices in the idea- Maley, M. E. 2010. Driving global innovation: Kellogg Com-
to-launch process and its governance. Research-Technology pany’s secret ingredient (People!). Presentation given at
Management 55(2): 43–54. Stage-Gate Innovation Summit, Clearwater, FL, November
Cooper, R. G., and Mills, M. 2005. Succeeding at new products 16–17.
the P&G way: A key element is using the “innovation dia- Morgan, J. 2002. Applying lean principles to product devel-
mond.” PDMA Visions 29(4): 9–13. opment. SAE in Manufacturing, February. http://www.sae.
Gehring, T. 2011. Sustaining an innovative culture at 3M. Pre- org/manufacturing/lean/column/leanfeb02.htm
sentation given at Stage-Gate Innovation Summit, Miami Spero, J. 2009. Lean in R&D. Presentation at IRI 2009 Six
FL, November 9–10. Sigma and DFSS in R&D Webinar Series, October.
Isaacson, W. 2011. Steve Jobs: The Exclusive Biography. New York: Takeuchi, H., and Nonaka, I. 1986. The new product develop-
Simon & Schuster. ment game. Harvard Business Review 64(1): 137–146.

VISIT the IRI BOOKSTORE! www.iriweb.org/bookstore

Access the best IRI has to offer in convenient electronic or print-on-demand versions. Browse white papers,
special offerings, and special reprint collections from RTM, selected and compiled to offer key perspectives
on central issues in the management of technological innovation.

SPECIAL PUBLICATIONS RTM REPRINT BOOKS

White Paper: Data Competitions The Art of Technology Management—79 articles

Driving breakthrough innovations Building an Innovation Culture *New*—47 articles
at breakneck speed . . . impact and Improving Product Development Processes—51 articles
limitations of competitions . . . using
Measuring and Improving the Performance and Return on R&D—
competition to enhance R&D.
54 articles
Free
Motivating, Appraising, Rewarding, and Retaining Engineers and
The Gender Diversity Playbook Scientists—53 articles
Best practices for fostering gender Promoting Breakthrough Innovation *Newly Updated*—54 articles
diversity in R&D and innovation R&D Funding & Valuation *New*—43 articles
management, gleaned from Succeeding as a New Manager/Leader—60 articles
the work of IRI’s Research-on-
Successful Portfolio Management *New*—32 articles
Research group on gender
diversity. Winning Concepts and Practices for Managing Industrial R&D—26 articles

$20.00 electronic only All reprint books $75 in hardcopy, $60 in electronic version

www.iriweb.org/bookstore

What’s Next? After Stage-Gate January—February 2014 | 31

Reproduced with permission of the copyright owner. Further reproduction prohibited without
permission.